A Few Brief Notes on DeepImpact, COIL, and a Conceptual Framework for Information Retrieval Techniques

TL;DR

This paper introduces a conceptual framework for information retrieval techniques based on representation types and compares recent methods, proposing a new simple extension called uniCOIL that achieves state-of-the-art sparse retrieval results.

Contribution

It presents a unifying framework for IR techniques, analyzes existing methods, and introduces uniCOIL, a novel extension that improves sparse retrieval performance.

Findings

uniCOIL achieves state-of-the-art results on MS MARCO

Framework clarifies relationships between IR techniques

Identifies unexplored opportunities in IR methods

Abstract

Recent developments in representational learning for information retrieval can be organized in a conceptual framework that establishes two pairs of contrasts: sparse vs. dense representations and unsupervised vs. learned representations. Sparse learned representations can further be decomposed into expansion and term weighting components. This framework allows us to understand the relationship between recently proposed techniques such as DPR, ANCE, DeepCT, DeepImpact, and COIL, and furthermore, gaps revealed by our analysis point to "low hanging fruit" in terms of techniques that have yet to be explored. We present a novel technique dubbed "uniCOIL", a simple extension of COIL that achieves to our knowledge the current state-of-the-art in sparse retrieval on the popular MS MARCO passage ranking dataset. Our implementation using the Anserini IR toolkit is built on the Lucene search library and thus fully compatible with standard inverted indexes.